Agricultural Engineering · Land & Water Resources Engineering · also Drainage Engineering, Management of Waterlogged and Problem Soils
Removing excess water and salt: surface and subsurface drain design, and reclaiming saline and sodic land.
In the exams: GATE AG Section 5 (drainage coefficient in 5 of 8, SAR and leaching in 3 of 8 papers 2019–26); ICAR AIEEA-PG Unit VII; ICAR AICE subjects 40 and 43; NCEES PE drainage; RPSC AAE
Agricultural drainage is the removal of excess water, and with it excess salt, from the land surface and the root zone so that crops can grow and machines can work. Too much water in the root zone starves roots of oxygen, delays sowing, cools the soil and, in dry climates, lets a shallow water table push salt up into the topsoil by capillary rise. Drainage engineering designs the systems that prevent this: open ditches and land shaping that carry off surface water, buried pipe drains that control the water table, wells that pump groundwater down, and the leaching and chemical amendments that reclaim saline and sodic land.
The subject sits on three foundations. The first is soil physics and groundwater flow: Darcy's law, hydraulic conductivity, drainable porosity and the steady and falling water-table equations of Hooghoudt, Ernst, Kirkham and Glover-Dumm. The second is hydraulics: Manning's equation for ditches and pipes, and filter rules for envelopes. The third is soil and water chemistry: electrical conductivity, the sodium adsorption ratio, exchangeable sodium, residual sodium carbonate, the leaching requirement and the gypsum requirement. A drainage engineer who knows only one of the three designs systems that fail: pipes that silt up, spacings that ignore a barrier layer, or leaching schemes that turn a saline-sodic soil into a dispersed, impermeable sodic one.
The same design problems appear in the humid Midwest of the United States and the Netherlands, where drainage is about rainfall and trafficability, and in the irrigated plains of India, Pakistan, Egypt, Iraq, Australia and California, where it is about salt. The equations are universal; the design criteria (how much water, how fast, to what depth) are local and are set by national agencies, research stations and experience. This manual says where values vary and does not pick one.
Examiners test the subject numerically. GATE AG repeatedly asks for the drainage coefficient, drainable porosity, drain discharge, SAR and the leaching requirement; ICAR PG and PhD papers add drain-spacing equations, envelope criteria, soil classification and reclamation; licensing exams such as the NCEES PE and the Philippine ABELE ask design questions in which the candidate must choose the criterion as well as compute. Work every formula in this manual with units written out: most wrong answers in this subject are unit slips between mm/day and m/day, hectares and square metres, or cmol/kg and meq/100 g.
Land is waterlogged when the root zone stays saturated, or nearly so, long enough to harm crops or hinder field work. The harm comes mainly from lack of air. Oxygen diffuses roughly ten thousand times more slowly through water than through air, so when the soil pores fill with water, the oxygen used by roots and soil microbes is not replaced. Within a day or two the soil becomes anaerobic: nitrate is lost by denitrification, manganese and iron are reduced to soluble and potentially toxic forms, sulphide and organic acids can form, and roots of most dryland crops stop growing or die back. Rice and some wetland plants tolerate saturation because they carry air to their roots through internal air spaces (aerenchyma).
Waterlogging is described either by the surface (ponding after rain or irrigation) or by the water table, the level at which the pore-water pressure equals atmospheric pressure. Many national agencies classify land by water-table depth; for example, an Indian working-group definition has treated land with a water table within about 2 m of the surface as waterlogged and 2 to 3 m as potentially waterlogged. Thresholds differ by country, crop and climate, so always use the definition the local agency uses. What matters agronomically is not a single depth but how long and how often the water table stays high during the growing season.
Natural causes are high or intense rainfall, flat land with poor natural outlets, slowly permeable subsoils or hardpans that hold water up (a perched water table), depressions that collect runoff, seepage from higher land, flooding by rivers or tides, and upward seepage from confined aquifers under artesian pressure.
Human causes are often larger in irrigated regions. Seepage from unlined canals and distributaries, over-irrigation and poor field levelling add water to the groundwater faster than it can leave. Before irrigation, a river basin's water table is in rough balance with natural recharge and outflow; after a canal system is built, recharge rises several-fold and the water table climbs, commonly by tens of centimetres a year, until it reaches the root zone. Roads, railways and embankments built across natural drainage lines block surface flow; drains that are never maintained silt up and weed over; rice grown in upland areas adds percolation to neighbouring fields. The historical lesson of irrigation development in the Indus, Nile, Tigris-Euphrates and Murray-Darling basins is the same: irrigation without drainage eventually produces waterlogging and salinity.
Crop effects: poor germination and stand, shallow rooting (which then exposes the crop to drought later in the season), nitrogen deficiency through denitrification and leaching, more root disease, and lower yield. Soil effects: wet soil warms slowly in spring because water has a high heat capacity and evaporation uses energy; wet soil has low bearing strength, so tillage and harvest are delayed and wheel traffic compacts and smears it; structure degrades. In arid and semi-arid climates the most serious long-term effect is secondary salinisation. When the water table is within reach of capillary rise, groundwater moves up to the surface, evaporates and leaves its salt behind. The depth at which this becomes serious (sometimes called the critical depth) depends on soil texture, groundwater salinity and evaporative demand; it is deeper in silty soils with strong capillary rise than in coarse sands.
Beyond the farm, waterlogged land breeds mosquitoes, damages village houses and roads, and reduces the area that can be cropped. Drainage therefore has benefits that are partly agronomic (timely operations, higher and more reliable yield, wider choice of crops) and partly public (health, infrastructure, salt control for a whole command area).
Before designing anything, the engineer finds out where the water comes from. The tools are simple. Observation wells (perforated along their length) show the water-table depth. Piezometers (open only at the bottom) show the hydraulic head at one depth; a nest of piezometers at different depths shows whether groundwater is moving downward (recharge area) or upward (discharge or artesian area). Upward head gradients point to deep seepage that needs relief drains or wells rather than ordinary field drains. Water-table hydrographs over a season, soil profiles showing mottling and grey (gleyed) colours, the position of slowly permeable layers, hydraulic conductivity measured by the auger-hole method, the depth to an impermeable barrier, and the salinity of groundwater and soil are the minimum data for a design. A water balance of the area (rainfall, irrigation, canal seepage, evapotranspiration, lateral inflow and outflow) tells whether the problem is local recharge, lateral seepage or regional groundwater rise, and so which type of drainage will work.
The drainage coefficient (DC) is the depth of water that a drainage system must remove from the drained area in 24 hours. It is written in mm/day, cm/day or, in the United States, inches per day. It converts a design criterion into a flow rate: the design discharge of a drain equals the drainage coefficient multiplied by the area it serves, Q = DC x A. Units are where students lose marks. One millimetre over one hectare is 10 cubic metres. So 1 mm/day from 1 ha is 10 m3/day, which is 10,000 L over 86,400 s, or about 0.116 L/s. A 10 mm/day coefficient on a 40 ha block therefore needs 10 x 40 x 10 = 4,000 m3/day, or about 46 L/s. In US units, one acre-inch per hour is almost exactly one cubic foot per second, so Q (ft3/s) is approximately DC (in/day) x A (acres) / 24.
For subsurface drains, the same quantity is usually called the design drain discharge or drainage rate, q, and is used in the spacing equations in m/day. A coefficient of 5 mm/day is q = 0.005 m/day. Mixing mm/day with K in m/day is the single most common error in spacing calculations.
In humid regions the drainage coefficient comes from rainfall: the excess of a design storm (a rainfall of chosen duration and return period, less what the soil stores or evaporates) that must leave within a set time. For subsurface drains in the humid US Midwest, coefficients of roughly 10 to 13 mm/day (3/8 to 1/2 inch per day) have long been used for field crops on mineral soils, with larger values for high-value crops, organic soils, or where surface inlets feed the pipes. These are regional practice values, not universal constants.
In arid irrigated regions rain is rare and the drainage coefficient comes from irrigation: the deep percolation from irrigation losses plus the water deliberately applied for leaching, plus any canal seepage reaching the area. If 100 mm is applied every 10 days and 25 per cent of it percolates below the root zone, the average recharge is 2.5 mm/day. Coefficients of about 1 to 4 mm/day are typical of irrigated arid lands, but the value must be computed from the local water balance.
For surface drainage the coefficient (or design runoff rate) is set from the design storm and the crop's tolerance to ponding; rice tolerates days of shallow submergence, while most dryland crops are damaged within one to two days.
A drainage criterion states the design discharge together with the water-table depth to be maintained. A steady-state criterion assumes recharge equals drain discharge and fixes the water-table height midway between drains: for example, a discharge of the order of 7 mm/day with the midpoint water table about 0.5 m below the surface has been used for arable land in the Netherlands. Such criteria suit humid climates with long wet periods.
A non-steady (falling water-table) criterion states how fast the water table must drop after a recharge event: for instance, lowering it from near the surface to about 0.3 m within one to two days, a type of criterion common in humid-area practice, or, in irrigated land, ensuring that the rise caused by each irrigation has drained away before the next. Non-steady criteria use the drainable porosity (also called specific yield or effective porosity), the volume of water released per unit volume of soil when the water table falls. A recharge R raises the water table by R divided by the drainable porosity: 30 mm of recharge into soil of drainable porosity 0.06 lifts it by 0.5 m. Drainable porosity is typically a few per cent in clays and 10 to 20 per cent or more in sands; an empirical relation attributed to van Beers relates it to hydraulic conductivity (porosity in per cent roughly equal to the square root of K in cm/day), but it should be measured where possible.
Two indices link water-table behaviour to yield. SEW30 (sum of excess water above 30 cm), introduced by Sieben, adds up, day by day during the growing season, the amount by which the water table stands above 30 cm depth; its unit is cm-days. A water table at 10 cm for three days and at 20 cm for two days gives (30 - 10) x 3 + (30 - 20) x 2 = 80 cm-days. Days when the water table is below 30 cm contribute nothing. The stress day index (SDI), developed by Hiler, multiplies a stress-day factor such as SEW30 for each growth stage by a crop susceptibility factor for that stage, because the same excess water hurts more at some stages than others. Simulation models such as DRAINMOD, developed by Skaggs, compute these indices over long weather records and so let designers compare drain spacings by expected yield and profit rather than by a single design storm.
Surface drainage removes water that ponds on the land or cannot infiltrate fast enough. It is the main solution on flat land with slowly permeable soils, in high-rainfall regions, in rice-growing areas, and wherever subsurface drains alone cannot take away intense storm runoff. It is also the collecting network (the main drains and outlets) for every subsurface system. A complete system has field drains or shaped surfaces within fields, collector drains along field boundaries, main drains that follow the natural low lines, and an outlet: a river, the sea (often through flap or tidal gates), or a pumping station where gravity outflow is impossible. Surface drainage is cheaper per hectare than pipe drainage but takes some land for ditches, needs regular cleaning, and does little for a high water table unless the ditches are deep. In humid flat lands it is often combined with subsurface drains, the surface system handling storm runoff and the pipes controlling the water table.
Random field drains are shallow ditches that connect isolated depressions to an outlet; they suit fields with scattered low spots. Parallel field drains (parallel ditches) are dug across very flat fields at regular spacing, with the land between them shaped to slope towards the ditches; spacing depends on soil, slope and the length of row the water must travel. Bedding is an older practice of ploughing land into crowned beds separated by dead furrows, which act as drains. Land grading shapes the whole field to a uniform slope so that water runs off without ponding, and land smoothing removes small humps and hollows; laser-controlled grading has made precise grading routine. Cross-slope drains and diversion ditches intercept runoff on sloping land. Shallow field ditches are given flat side slopes so that machines can cross them.
Choice of system depends on slope, soil depth available for cutting (grading must not expose infertile subsoil), crops, machinery and cost.
The design flow comes from a design storm. For small catchments the rational method is used: Q = C I A / 360, with Q in m3/s, I the rainfall intensity in mm/h for a duration equal to the time of concentration, A in hectares and C the runoff coefficient. For larger, flat agricultural catchments, the peak does not scale linearly with area, because flat land releases water slowly. The Cypress Creek formula, used in United States practice, has the form Q = C A^(5/6), where the coefficient C depends on the design rainfall excess and the land; this form is found in NRCS drainage guidance and ASABE engineering practice for surface drainage in humid areas. Many countries use their own drainage-module rule: a specific discharge in L/s per hectare or a depth to be removed within a stated number of days, chosen from local rainfall statistics and crop tolerance.
Open drains are sized with Manning's equation: Q = (1/n) A R^(2/3) S^(1/2), where A is flow area, R = A/P the hydraulic radius (P the wetted perimeter), S the bed slope and n the roughness coefficient. For a trapezoidal section with bed width b, depth y and side slope z horizontal to 1 vertical, A = by + zy^2 and P = b + 2y sqrt(1 + z^2). Earth drains in fair condition have n of about 0.025 to 0.035; weedy drains are much rougher, which is why maintenance is part of design. Velocity must be low enough not to erode the banks and high enough to avoid silting; permissible values depend on soil and are tabulated in design manuals. Side slopes are set by soil stability, flatter in sands and unstable soils. Drain depth is set by the need to receive outflow from subsurface drains and field drains, so the water level in the collector at design flow must lie below the outlets of the laterals; a freeboard is added above the design water level. Design is usually done upstream to downstream, adding catchment area at each junction and checking capacity, velocity and water level at each reach.
Parallel buried drains (laterals) at spacing L and depth below the surface lower the water table to a curved shape: highest midway between drains and lowest at the drains. Under steady state, recharge q (m/day) enters uniformly across the surface and the same amount leaves through the drains. The design unknown is L such that the water-table height midway between the drains, h, measured above drain level, does not exceed the criterion. If the criterion is a water table at 0.5 m depth and the drains are at 1.2 m depth, then h = 0.7 m. Steady-state equations rest on the Dupuit-Forchheimer assumptions: flow is horizontal and the hydraulic gradient equals the slope of the water table. These are good far from the drains but fail near them, where flow lines converge radially towards a small pipe. Steady-state equations also assume a homogeneous or simply layered soil, uniform recharge and drains running at free outfall.
If an impermeable layer lies at depth D below drain level and the drains are imagined as full-depth ditches, Dupuit-Forchheimer gives Donnan's equation, L^2 = 4K(H^2 - D^2)/q, where H = D + h is the midpoint water-table height above the barrier. Expanding, L^2 = (8KDh + 4Kh^2)/q.
Hooghoudt (1940) recognised that real drains are pipes, not ditches reaching the barrier, and that the converging radial flow near a pipe loses extra head. He kept the same form but replaced D by an equivalent depth d, smaller than D, which adds the right amount of resistance: L^2 = (8 K2 d h + 4 K1 h^2)/q. Here K1 is the hydraulic conductivity above drain level and K2 below drain level; for a uniform soil K1 = K2 = K. The first term is flow below drain level; the second is flow above it. When drains rest on the impermeable layer, d = 0 and L^2 = 4Kh^2/q.
The equivalent depth depends on D, L and the drain radius r, and Hooghoudt published tables. A widely used approximation (Moody, 1966) is: for D/L up to 0.3, d = D / [1 + (D/L)(8/pi x ln(D/r) - 3.4)]; for D/L above 0.3, d = pi L / [8 (ln(L/r) - 1.15)]. Because d depends on L, the solution is iterative: guess L, find d, compute L, repeat until it stops changing, usually within a few cycles. Note that once D is more than about a quarter of L, making the barrier deeper hardly changes d, because flow does not penetrate much deeper than that.
Ernst (1956, 1962) split the total head loss h into three parts, each with its own resistance, which makes the method convenient for layered soils and for drains lying in any layer: h = hv + hh + hr = q Dv/Kv + q L^2/(8 Sum(KD)) + (q L/(pi Kr)) ln(a Dr/u). The vertical term is the head lost in flowing downward through the layer of thickness Dv above the drains with vertical conductivity Kv; it is usually small. The horizontal term is the Dupuit flow through the layers, where Sum(KD) is the transmissivity of the layers taking part in horizontal flow. The radial term represents convergence to the drain: Kr is the conductivity of the layer in which radial flow occurs, Dr the thickness of that layer below the drain (limited in practice to about a quarter of L), u the wetted perimeter of the drain (for a pipe with an envelope, the envelope's outer perimeter), and a a geometry factor that is 1 in a homogeneous profile and depends on the conductivity contrast when the drain lies in a layer above a more or less permeable one. Because L appears in two terms, the equation is solved as a quadratic in L or by trial.
The steps are: (1) set the criterion (q and water-table depth); (2) choose drain depth from outlet levels, machinery and, in saline areas, the need to keep the water table below the depth of serious capillary rise; (3) measure K at many points, usually by the auger-hole method, and find the depth to the barrier; (4) choose drain radius and envelope; (5) compute L with Hooghoudt for fairly uniform profiles or Ernst for layered ones; (6) round down to a practical spacing and check discharge per lateral, Q = q x L x lateral length. Spacing varies roughly as the square root of K and of q inversely, so an error of a factor of four in K changes L by a factor of two. K varies over orders of magnitude in the field, which is why the geometric mean or a careful representative value of many measurements is used, and why pilot areas are worth their cost. Deeper drains allow wider spacing but cost more per metre and need a deeper outlet; in arid saline land, deeper drains are chosen mainly to control capillary salinisation, while very deep drains also collect more saline regional groundwater that must then be disposed of.
Hooghoudt's and Ernst's equations are approximations built on horizontal-flow assumptions plus a radial correction. Kirkham (from 1958 onwards) solved the flow problem more rigorously by potential theory: he solved Laplace's equation for flow to an array of parallel drains above an impermeable layer, without assuming horizontal flow. His solutions are written with a geometry function (often denoted F) that depends on drain radius, drain depth, spacing and depth to the barrier, and is given as a series or in charts.
Kirkham's best-known case is ponded water: water standing on the soil surface and flowing down to drains. This is exactly the situation during leaching reclamation of saline soil and in flooded rice, and the drainage model DRAINMOD uses Kirkham's ponded-water equation for the drain flux whenever the surface is ponded, switching to Hooghoudt's equation when the water table falls below the surface. A practical consequence of Kirkham's solution matters to reclamation: under ponding, water enters the soil much faster near the drain lines than midway between them, and flow paths to the midpoint are longest. Leaching midway between drains is therefore slowest, and the salt left there governs how long reclamation takes. Kirkham's solutions also serve as the reference against which the approximate spacing formulas were checked.
Rainfall and irrigation arrive in pulses, not as a steady trickle. After each pulse the water table rises suddenly and then falls as the drains discharge. In irrigated arid land, the United States Bureau of Reclamation developed a design approach around this cycle: each irrigation raises the water table by the deep percolation divided by the drainable porosity, the drains lower it before the next irrigation, and over a season the system reaches a dynamic equilibrium in which the rise from each irrigation equals the fall between irrigations. The spacing is chosen so that the equilibrium maximum water table stays below the root zone. The falling water-table equation used for this is the Glover-Dumm equation. Because the calculation follows the water table through a series of irrigations, it reflects the real schedule, the drainable porosity and the time available between irrigations, which a single steady-state discharge cannot capture. It is the standard approach in irrigated arid land.
Glover solved the one-dimensional linearised (Boussinesq) equation for a water table falling between parallel drains. Dumm assumed the initial water-table shape to be a fourth-degree parabola, which is closer to what is seen after irrigation, and obtained, for the midpoint height ht after time t starting from h0: ht = 1.16 h0 exp(-alpha t), with alpha = pi^2 K D/(mu L^2), the reaction factor (also called the reservoir coefficient), in 1/day. Rearranged for spacing: L^2 = pi^2 K D t / [mu ln(1.16 h0/ht)]. Here K is hydraulic conductivity, mu the drainable porosity, and D the average thickness of the flow region; in US Bureau of Reclamation practice D is commonly taken as the equivalent depth plus half the initial water-table height, d + h0/2, while some texts use d alone, so state the convention you use. If the initial water table is flat (horizontal) rather than parabolic, the factor 1.16 is replaced by 4/pi, about 1.27 (Glover's original case).
The equation shows that spacing grows with the square root of K, D and the time allowed, and shrinks with the square root of drainable porosity: a soil that releases more water per metre of water-table fall needs closer drains to achieve the same fall in the same time. The reaction factor alpha is a useful single number: a large alpha (say above 0.1 per day) means a fast-responding system, a small one a sluggish system.
Early subsurface drains were short clay or concrete tiles laid end to end, with water entering at the joints; hence the American word 'tile drainage' for any buried drain. Since the 1960s and 1970s corrugated plastic pipe (PVC or polyethylene) has largely replaced tiles: it is light, comes in long coils, is laid by machine, and admits water through slots cut in the valleys of the corrugations. Rigid smooth-wall plastic or concrete pipes are used for larger collectors.
Layouts are singular, where each lateral discharges directly into an open collector ditch, or composite, where laterals join a buried collector pipe that discharges to an open drain or sump. Singular systems are easy to inspect and maintain; composite systems take no land for ditches and suit large mechanised fields but need inspection chambers (manholes) at junctions. Laterals are laid on a uniform grade; very flat grades risk sediment deposition, and grades are commonly kept above about 0.1 per cent where the land allows, though standards differ. Installation depth in humid regions is often around 0.8 to 1.5 m; in arid saline regions deeper, often 1.5 m or more, to control capillary rise. These are ranges from practice, not rules.
The required capacity of a pipe at its outlet is Q = q x A, where A is the area drained by that pipe. Pipe diameter is chosen with Manning's equation for full flow: Q = (1/n)(pi D^2/4)(D/4)^(2/3) S^(1/2), which gives D in terms of Q, n and S. Roughness is lowest for smooth pipes (n of the order of 0.011) and higher for corrugated pipes (of the order of 0.015 to 0.02, rising with diameter because the corrugations are larger); use the value in the standard you design to. In a lateral, flow increases from zero at the upper end to the full design flow at the outlet, so a lateral sized on the outlet flow is conservative; design charts in NRCS and ASABE practice account for this. Laterals are usually of a minimum practical diameter set by installation and cleaning rather than by hydraulics. Collectors, which carry the flow of many laterals, are sized reach by reach as area accumulates.
A drain envelope is porous material placed around the pipe. It has three functions: a filter function (retaining soil particles so they do not enter and clog the pipe), a hydraulic function (a permeable zone that reduces entrance resistance and in effect enlarges the drain's radius), and a bedding function (support for the pipe). Envelopes are needed in unstable soils, notably fine sands and silts with little clay, which flow into drains; well-structured clay soils are often stable and may need none. FAO Irrigation and Drainage Paper 60 gives the soil criteria.
Granular envelopes (graded sand and gravel) are designed with filter rules of the Terzaghi type: the envelope's D15 must be small enough relative to the soil's D85 to retain the soil (a ratio of about 4 to 5 is the classic limit), and large enough relative to the soil's D15 to be more permeable than the soil. Organic envelopes (straw, coconut fibre, peat) are cheap but decompose; they are used where they last long enough for the soil around the drain to stabilise. Synthetic envelopes (geotextiles and pre-wrapped loose fibre) are specified by their characteristic opening size, O90, compared with the soil's d90; the permitted ratio differs between standards.
Drains are laid by trenching machines, which dig a trench and place pipe and envelope, or by trenchless ploughs, which lift the soil and draw the pipe in behind a blade; both use laser grade control. Installing in wet soil smears the trench wall and seals the soil next to the drain, one of the commonest causes of poor performance. Clogging comes from sediment (soil entering through inadequate envelopes), from ochre (iron oxide deposits formed when dissolved ferrous iron in groundwater is oxidised at the drain), from roots of trees and perennial crops growing into pipes, and from carbonate precipitation. Ochre is most severe in the first years after installation in soils rich in reduced iron. Pipes are cleaned by high-pressure jetting through the outlet or manholes. Outlets need protection against erosion and animal entry, and a free outfall above the collector water level.
Mole drains are unlined channels formed in clay subsoil by a bullet-shaped mole plough, commonly around 0.5 to 0.7 m deep and a few metres apart, often feeding into permanent pipe collectors. They are cheap and effective in stable clays but last only several years and must be redrawn.
Vertical drainage lowers the water table by pumping groundwater from wells. It works where a permeable aquifer is connected to the waterlogged topsoil, so that pumping from depth draws the water table down over a wide area. Its great advantage arises when the pumped water is fresh enough to irrigate with: the drainage water becomes an irrigation supply, and drainage and irrigation become one conjunctive system. The largest example is Pakistan's Salinity Control and Reclamation Projects (SCARPs), begun in the early 1960s, which installed thousands of public tubewells in the Indus plains. They lowered water tables over large areas, but running costs, power supply, maintenance and, where groundwater was saline, disposal of pumped water became serious problems, and many were later transferred to or replaced by private wells.
Design uses well hydraulics (see the groundwater manual): the steady pumping rate needed to hold the water table stable equals the recharge over the area served, Q = R x A, and wells are spaced so their cones of depression overlap enough to keep the water table below the target everywhere. Where fresh water floats on saline groundwater, deep pumping draws up the saline water (upconing). Skimming wells, shallow wells or multiple shallow strainers pumping at low rates, take only the upper fresh layer.
Bio-drainage uses vegetation with high water use, usually trees such as eucalyptus and poplar, or deep-rooted perennial grasses, to take groundwater out through transpiration. Plantations are placed along seeping canals, in strips across fields or in blocks in waterlogged areas. Advantages: no pumping energy, no drainage effluent to dispose of, low cost, and products (timber, fuel, fodder). Limitations: it acts slowly, it takes land out of crops, water use falls in winter and in young plantations, and, where groundwater is saline, the salt the trees exclude accumulates in their root zone and eventually reduces their growth, so bio-drainage alone is not a sustainable salt outlet. The land fraction needed follows from a water balance: if a plantation can take up a known extra amount of groundwater per day beyond rainfall, the planted fraction must be large enough for that uptake to balance recharge over the whole area. FAO's synthesis on bio-drainage (Heuperman, Kapoor and Denecke, 2002) summarises field experience.
Dry drainage leaves part of an irrigated area fallow so that it acts as an evaporation sink; groundwater flows laterally from irrigated land to the fallow land, where salt accumulates. It needs no outlet but sacrifices land. Controlled drainage (drainage water management) fits the outlet with a structure that raises the water level in the drains when drainage is not needed, conserving water and reducing nitrate losses, and lowers it before field operations. Evaporation ponds receive saline drainage water where no outlet exists, but they need large areas and can concentrate toxic elements; selenium in drainage evaporation ponds in California's San Joaquin Valley caused well-documented harm to waterbirds in the 1980s.
Drainage water can be reused: blended with fresh water to a salinity the crop tolerates, or used cyclically, with fresh water for sensitive stages and saline water for tolerant crops or stages. Each reuse concentrates salt further, so the final effluent must still leave the basin. FAO Irrigation and Drainage Paper 61 treats drainage water management in arid areas.
An interceptor drain is a single drain, usually laid across the slope, that cuts off groundwater flowing laterally into an area from higher ground or from a canal. Typical uses are hillside seepage, where water moving down a slope over a slowly permeable layer emerges at a break of slope, and seepage from a canal or reservoir running alongside farmland. The drain is placed on the upslope side of the wet area, at right angles to the direction of groundwater flow. It works best when it is deep enough to reach or cut into the impermeable layer, so that it intercepts the whole flow; a drain that does not reach the barrier lets part of the flow pass beneath it, and its effect downslope dies out with distance. The inflow is estimated by Darcy's law: flow per metre length = K x i x b, where i is the hydraulic gradient (often close to the land slope for hillside seepage) and b the thickness of the saturated flow zone. For example, K = 2 m/day, i = 0.05 and b = 3 m give 0.3 m3/day per metre, or 150 m3/day for a 500 m drain. A single interceptor rarely drains a wide area on its own; beyond its zone of influence, relief drains are needed. Canal-side interceptors are often combined with canal lining or bio-drainage strips.
Relief drains are a pattern of drains, usually parallel laterals, installed within the wet area to lower a water table that is fed from above (rain, irrigation) or from below (upward seepage). They are what the spacing equations of chapters 4 and 5 design. The distinction is about the water source: interceptors deal with water arriving laterally from outside the area; relief drains deal with water arriving over the area itself.
Where the problem is upward artesian pressure from a confined aquifer beneath a less permeable layer, ordinary shallow drains may be overwhelmed. Relief wells, which penetrate the confining layer and let water rise freely to a drain or pump, reduce the pressure directly. Relief wells are also used at the landward toe of levees and dams to control uplift pressure. Diagnosis by piezometer nests is essential here: if deeper piezometers show heads above the water table, water is coming up, and drains designed only for surface recharge will be spaced too wide.
Soluble salts, mainly chlorides, sulphates and bicarbonates of sodium, calcium and magnesium, come from weathering of rock, from irrigation water, from saline groundwater rising by capillarity, from seawater in coastal lands and from fossil marine deposits. Every irrigation adds salt: water of electrical conductivity 1 dS/m carries roughly 640 mg/L of dissolved salts, so 1,000 mm of it over a hectare (10,000 m3) adds about 6.4 tonnes of salt. Crops take up very little of it, so the salt stays in the soil unless water moves it downward.
Soil salinity is expressed as ECe, the electrical conductivity of the saturation-paste extract, in dS/m (numerically equal to mmho/cm). Rapid methods measure extracts at fixed soil:water ratios such as 1:2 or 1:5, which must be converted to ECe with soil-specific factors, and electromagnetic induction sensors survey whole fields. Rough conversions for water are: total dissolved solids (mg/L) about 640 x EC (dS/m), the factor ranging roughly from 550 to 700 with ionic composition, and total cations (meq/L) about 10 x EC (dS/m).
Salt harms crops by the osmotic effect (salt lowers soil-water potential, so plants must spend energy to take water up; the effect resembles drought), by specific-ion toxicity (sodium, chloride and boron, especially to trees and vines), and by nutritional imbalance. Maas and Hoffman (1977) described relative yield with a threshold-slope model: Y (%) = 100 - b(ECe - a) for ECe above the threshold a, where b is the per cent yield loss per dS/m. Crop thresholds and slopes are tabulated in FAO Irrigation and Drainage Papers 29 and 48 and in USDA Salinity Laboratory publications; they depend on climate, variety and management and are used as guides. Tolerance varies widely: barley, cotton and sugar beet are tolerant; many beans, fruit trees and vegetables are sensitive. Salinity also often matters most at germination and early growth. Sprinkler irrigation with saline water can additionally injure leaves by direct absorption of sodium and chloride.
To stop salt building up, more water must pass through the root zone than the crop uses. The leaching fraction (LF) is the fraction of applied water that drains below the root zone: LF = Ddw/Diw, where Ddw is drainage depth and Diw irrigation depth. At steady state, salt entering with irrigation equals salt leaving in drainage, so Diw x ECiw = Ddw x ECdw and LF = ECiw/ECdw. The leaching requirement (LR) is the minimum leaching fraction that keeps root-zone salinity acceptable. The US Salinity Laboratory (Handbook 60, 1954) wrote LR = ECiw/ECdw, where ECdw is the highest drainage-water salinity allowed. Rhoades's later formula, adopted in FAO Irrigation and Drainage Paper 29 (revision 1), links it to crop tolerance: LR = ECw/(5 ECe - ECw), where ECe is the average root-zone soil salinity the crop tolerates (often the threshold for 100 per cent yield).
The total water to apply is Diw = ETc/(1 - LR) when leaching is supplied with each irrigation, so the extra depth is ETc x LR/(1 - LR). In fine-textured or cracking soils some water bypasses the soil matrix through large pores and leaches little; a leaching efficiency coefficient (less than 1) is then applied by dividing LR by it. The leaching water must go somewhere: without drainage, leaching raises the water table and returns salt to the surface. Leaching requirement and drainage coefficient are two views of the same design.
The same balance applies to a whole irrigation scheme. Salt imported with canal water, plus salt mobilised from the subsoil, must be exported through drains or it accumulates. Drain effluent salinity therefore often exceeds the root-zone target, particularly in the first years after drainage when stored salt is being removed. The long-term sustainability test for an irrigated area is whether the salt balance, measured at the drainage outlet, is neutral or negative. A positive balance (more salt in than out) means salt is being stored somewhere in the soil or groundwater, and will appear later as rising soil salinity or more saline groundwater. Measuring the balance needs flow and EC records at canal heads and drain outlets over several years, because stored salt moves slowly and seasons differ. Such records also show whether drains are working at all. Where the balance is negative, the area is being desalinised, which is the aim of reclamation.
The US Salinity Laboratory classification (Handbook 60, 1954) remains the one examined everywhere. Saline soils have ECe above 4 dS/m, exchangeable sodium percentage (ESP) below 15 and pH usually below 8.5; they often carry white salt crusts ('white alkali') and have good or even flocculated structure. Sodic (alkali) soils have ESP above 15, ECe below 4 dS/m and pH commonly between 8.5 and 10; the organic matter disperses and darkens the surface ('black alkali'), clay disperses, and the soil becomes hard when dry and nearly impermeable when wet. Saline-sodic soils have ECe above 4 dS/m and ESP above 15; the high salt concentration keeps the clay flocculated and pH is usually 8.5 or below. ESP = exchangeable sodium / cation exchange capacity x 100. Some classifications use SAR of the saturation extract above 13 in place of ESP 15, and some soil scientists now argue for lower salinity thresholds such as 2 dS/m; say which you use. In the Indo-Gangetic plains, sodic lands are widely called usar.
Saline soils are reclaimed by leaching with good-quality water, provided the water table can be kept deep, which usually means installing drainage first. The amount of water depends on the initial and target salinity, the soil and the method. An empirical relation summarised by Hoffman states that (C/C0) x (Dw/Ds) is roughly constant, where C/C0 is the fraction of the initial salt concentration remaining, Dw the depth of leaching water infiltrated and Ds the depth of soil to be reclaimed; the constant is about 0.3 for continuous ponding in many soils and about 0.1 for intermittent ponding or sprinkling. So intermittent application removes the same salt with roughly a third of the water, because more of the water moves slowly through the fine pores where the salt is. The constants are soil-specific and should be checked in a trial. Leaching in stages, with a salt-tolerant first crop, and land levelling so that water infiltrates evenly, are standard practice.
Sodic soils cannot be reclaimed by water alone. Exchangeable sodium must be replaced by calcium, and the displaced sodium (as sodium sulphate or chloride) leached out. Calcium sources: gypsum (calcium sulphate dihydrate), the most used, cheap and safe to handle; phosphogypsum, a by-product of phosphoric acid manufacture; calcium chloride. Acid-forming amendments (elemental sulphur, sulphuric acid, iron sulphate, pyrite) contain no calcium; they work only in calcareous soils, where the acid they produce dissolves soil calcium carbonate and releases calcium. Organic matter, green manures such as sesbania (dhaincha), and rice as the first crop help, because rice tolerates sodicity and the ponded water aids leaching; salt-tolerant grasses such as kallar grass (Leptochloa fusca) are used in biological reclamation.
Saline-sodic soils must not be leached with good water before the amendment is applied: removing the salt first lets the clay disperse, and the soil becomes a sodic soil of very low permeability. Apply the calcium source first, then leach.
The gypsum requirement (GR) is the amount of gypsum needed to lower ESP from its initial value to a target in a given depth of soil. Chemically, GR (cmol/kg or meq/100 g) = (ESPi - ESPf) x CEC/100. Convert to mass: gypsum (CaSO4.2H2O, molar mass about 172 g) has an equivalent weight of about 86 g, so 1 meq needs 86 mg of gypsum. Soil mass per hectare = 10,000 m2 x depth (m) x bulk density (t/m3). Example: ESP 40 to 10, CEC 15 cmol/kg, 0.15 m depth, bulk density 1.5 t/m3. GR = 30/100 x 15 = 4.5 cmol/kg = 45 meq/kg. Soil mass = 2,250 t/ha = 2.25 million kg. Gypsum = 45 x 2.25 x 10^6 x 86 mg, about 8.7 t/ha of pure gypsum. Divide by purity for commercial gypsum, and many practitioners multiply by a factor above 1 because the Ca-Na exchange is not complete in the field. The laboratory gypsum requirement test of Schoonover (in Handbook 60) measures it directly by equilibrating soil with saturated gypsum solution.
Other amendments are compared on a chemical-equivalent basis: 1 t of sulphur (32 g/mol, producing one sulphate) is equivalent to about 172/32, or 5.4 t of gypsum; 1 t of gypsum is equivalent to about 98/172, or 0.57 t of sulphuric acid. Gypsum is usually applied to the surface 0.10 to 0.15 m, mixed in, and followed by ponding and leaching.
Water quality for irrigation is judged on total salinity, sodium hazard, bicarbonate (carbonate) hazard, and specific toxic ions such as boron and chloride. Ion concentrations for these indices must be in milliequivalents per litre (meq/L). Convert mg/L to meq/L by dividing by the equivalent weight (molar mass divided by charge): sodium 23, calcium about 20, magnesium about 12.2, potassium 39, chloride 35.5, sulphate 48, bicarbonate 61 and carbonate 30. For example, 460 mg/L of sodium is 20 meq/L, and 200 mg/L of calcium is 10 meq/L. Forgetting this conversion, or using mmol/L instead of meq/L for calcium and magnesium, is a common source of wrong SAR values.
Salinity is measured as ECw in dS/m (or microsiemens per centimetre, uS/cm; 1 dS/m = 1,000 uS/cm). The USDA Salinity Laboratory classes are C1 (low, below 250 uS/cm), C2 (medium, 250 to 750), C3 (high, 750 to 2,250) and C4 (very high, above 2,250). FAO Irrigation and Drainage Paper 29 (Ayers and Westcot, 1985) gives degrees of restriction on use: none below 0.7 dS/m, slight to moderate from 0.7 to 3.0, and severe above 3.0 dS/m.
The sodium adsorption ratio expresses the sodium hazard: SAR = Na / sqrt((Ca + Mg)/2), with all concentrations in meq/L; its unit is (meq/L)^0.5 but it is usually quoted as a number. It predicts the exchangeable sodium the soil will reach in equilibrium with the water. Handbook 60 gives the empirical relation ESR = -0.0126 + 0.01475 SAR, where ESR is the exchangeable sodium ratio (exchangeable Na / other exchangeable cations), and ESP = 100 ESR/(1 + ESR). In the USSL diagram, water is placed in sodium classes S1 to S4 by SAR and salinity together; the SAR boundaries between classes fall as salinity rises. The infiltration hazard depends on SAR and ECw jointly: at a given SAR, low-salinity water disperses soil more than saline water, which is why very pure water (rain, snowmelt, some canal water) can cause crusting and poor infiltration. FAO 29 therefore reads the infiltration hazard from SAR and ECw together. Refinements such as the adjusted SAR account for calcium precipitation or dissolution in the soil when the water contains bicarbonate.
Water rich in bicarbonate and carbonate relative to calcium and magnesium precipitates calcium and magnesium as carbonates in the soil as it concentrates, raising the effective sodium proportion and the pH. Eaton's residual sodium carbonate expresses this: RSC = (CO3 + HCO3) - (Ca + Mg), all in meq/L. Widely used USDA guidance classes water with RSC below 1.25 meq/L as safe, 1.25 to 2.5 as marginal and above 2.5 as unsuitable; several countries, including India, use their own class limits, so quote the scheme you use. High-RSC groundwater is common in parts of the Indo-Gangetic plain and in other alluvial aquifers. It can be used with care by adding gypsum to the soil or to the water (for example gypsum beds in the channel), by blending or alternating with canal water, and by growing tolerant crops, with regular monitoring of soil pH and ESP. The gypsum dose for water is calculated from the RSC to be neutralised and the volume of water applied.
A water-quality assessment for drainage and irrigation design reports: ECw, the major cations and anions in meq/L with an ion-balance check (sum of cations roughly equal to sum of anions), SAR, RSC, boron and chloride, and a classification by the scheme used locally. An ion balance that is off by more than a few per cent signals an analytical error to be resolved before any index is computed. The designer then decides the leaching requirement, whether amendments are needed, which crops suit the water, and whether drainage water can be reused. The same indices apply to drainage effluent planned for reuse, where salinity and sodium are higher than in the original irrigation water. Water quality also changes over time: pumping can draw more saline or high-bicarbonate water into wells, so samples should be taken at least seasonally and compared with earlier results. Reports should state the sampling date, source and the laboratory method, because EC depends on temperature and is normally reported at 25 degrees Celsius.
Learn the drainage coefficient conversion (1 mm/day over 1 ha = 10 m3/day) until it is automatic, and practise Hooghoudt with the equivalent-depth iteration by hand at least twice. Memorise the Handbook 60 classification limits for saline, sodic and saline-sodic soils, the SAR and RSC formulas with meq/L units, the two leaching requirement formulas and when each is used, and the gypsum-requirement chain from ESP to tonnes per hectare. In exams, write units on every line: most drainage marks are lost to mm versus m and cmol/kg versus t/ha slips.
A field engineer five years in designs drainage for a block or a pilot area. The daily decisions are: which criterion to use for this crop and climate, how many auger-hole measurements of K are enough, whether the barrier depth is real, which envelope the soil needs, and whether the collector can take the lateral outflow at design water level. In reclamation work the engineer computes gypsum and leaching water, sequences amendment before leaching on saline-sodic land, and checks by resampling ECe and ESP after each season. Supervising installation (grade, avoiding smearing in wet soil, envelope placement) matters as much as the calculation.
A senior engineer with 25 years plans drainage for whole command areas: choosing between surface, subsurface, vertical and bio-drainage by water balance and economics; deciding where saline effluent goes; setting maintenance organisation and cost recovery with farmers' groups; and judging when a pilot shows a design criterion is wrong. They know why many drainage projects failed (no outlet, no maintenance, poor installation, underestimating seepage) and write specifications, tender documents and monitoring programmes accordingly, including salt-balance monitoring at the outlet. They also negotiate with irrigation agencies over canal lining and water allowances, because seepage control upstream can be cheaper than drains downstream.
A researcher with 25 years' experience works on the questions design practice cannot yet answer well: scaling point measurements of hydraulic conductivity to field design values, transient simulation of water table, yield and salt over decades (DRAINMOD-type and variably saturated models), controlled drainage for nitrate and phosphorus loss, drainage effluent reuse and disposal, sodic water management under changing groundwater quality, and bio-drainage sustainability. They review national design criteria against long-term data, and they train engineers to test a design in pilot areas before scaling up.
This manual gives the equations and the reasoning but not the full design tables: Hooghoudt's equivalent-depth tables, Kirkham's F-function series, crop salt-tolerance thresholds and slopes, NRCS drainage-coefficient tables and Cypress Creek coefficients, envelope O90/d90 limits and pipe roughness values must be taken from the named standards. It does not yet cover drainage of organic (peat) soils and subsidence, drainage of rice paddies in detail (including Japanese and Southeast Asian paddy practice), tidal and polder pumping-station design, drainage economics and cost-benefit analysis, nitrate and phosphorus losses and wetland treatment of drainage water beyond a brief mention, or the auger-hole and inverse-auger-hole equations for measuring hydraulic conductivity. National waterlogging definitions and water-quality class limits used by India, Pakistan, Egypt and other countries are mentioned only in outline.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.